Career Development

Topic: Artificial Intelligence - Deep Learning

Introduction

What is Deep Learning?

• Deep learning (subset of machine learning) is a specialized form of artificial intelligence that imitates the operations of the human brain.

• It uses artificial neural networks that are specifically made to be able to learn and think as humans do.

  • The neural networks allow images, sounds, and texts to be processed and used as information.

• It has been used for image classification, language translation, and speech recognition.

History

1943

• Walter Pitts and Warren McCulloch created the first computer model with artificial neural networks was made using algorithms and mathematics to form what is called the “threshold logic”.

• Threshold logic was created to mimic the thought process of the human brain.

1960

• Henry J. Kelley formed the basics of the Back Propagation Model was developed.

• The Back Propagation Model was used for neural net training

1965

• The earliest effort found in forming Deep Learning algorithms by using polynomial activation functions was done by Alexey Grigoryevich Ivakhnenko and Valentin Grigorʹevich Lapa.

1979

• Kunihiko Fukushima

• He developed Neocognitron, an artificial neural network using a hierarchical, multilayered design.

  • The design allowed the computer to learn and recognize visual patterns.

Present

• Companies have began to use deep learning AIs to find useful, hidden informations that could be helpful for their needs.

• Computers, machines, and programs today have become sharper and a lot more developed, almost reaching human-level capabilities.

Examples of Deep Learning Models & AIs:

• Computer Vision

• Automated Translations

• Bots (based on deep learning)

• Autonomous vehicles

Other Examples

• Apple: Used to improve Siri’s voice

• Alphabet Inc / Waymo: Used to create self-driving cars

• Baidu: Uses deep learning to clone a voice

Future

Future of Technology

• It is predicted that within the next 5 to 10 years, Deep Learning AI would most likely be accessible to everyone through computing platforms.

• As technology is constantly improving, it may be beneficial in the development of deep learning AI, providing more use and allowing deep learning AIs to mimic the functions and working process of the human brain.

How It Affects the Design Industry (Opinion)

• Deep learning AIs would most likely come in the form of applications, machines, and gadgets. With this in mind, the design industry could be positively affected as the said forms require appealing, effective, and efficient designs that are unique, fits the product’s image, and suitable to the taste of the consumers.

• Promotion requires unique and eye-catching ads. Ad designs and typography needs to be approach methodically in a way that would cause consumers to be interested in the products.

---

Feedback

- Add conclusion

- Add more case study for the use of deep learning in the present day

---

Present

Case Study: Microsoft - Tay

Microsoft had released 2 deep learning AI bots, known as Tay and Zo. Tay — Microsoft’s first English language Twitter bot released on March 23, 2016 — was originally created as an experiment in “conversational understanding”. However, just a few hours after its release, Tay began exhibiting unacceptable behaviours due to people tweeting improper contents, such as misogyny and racism. As a result, Tay was then shut down just 16 hours after its release.

Conclusion and Summary

• Deep learning have been under development since the 1900s through the different findings, experiments, and upgrades. As a result, deep learning has become one of the essential core in developing AIs that could be another step to the future development of technology.

• Design may also be important in the development of these technology as it require an effective, appealing, and efficient design. This may affect the design industry positively, however, due to the increasing need for design, competition may also increase.

---